1. Introduction
A central concern of present-day climate change is the melting of the West Antarctic ice sheet (WAIS) and its potential to contribute roughly 50 cm to global sea level rise over the next century if emissions remain unchecked (Bamber et al. 2019; DeConto et al. 2021; Edwards et al. 2021). The present-day negative mass balance for the WAIS is driven by intrusions of relatively warm Circumpolar Deep Water onto the continental shelf, inducing basal melt at the glaciers’ grounding lines (Goddard et al. 2017; Jenkins et al. 2016; Rignot and Jacobs 2002). While the WAIS precipitation budget is insufficient to offset the basal melt from 1992 to 2018 (Frieler et al. 2015; Lenaerts et al. 2016; Shepherd et al. 2018), Paolo et al. (2018) show that increased precipitation over the WAIS due to warming may offset much of the overall mass loss. Therefore, thoroughly understanding seasonal to annual drivers of WAIS precipitation variability, in addition to surface temperature variability, is essential to diagnosing the ice sheet’s total mass in the past, present, and future (Smith et al. 2020).
Seasonal to annual precipitation and surface air temperature variability at the WAIS is largely controlled by the position and depth of the Amundsen Sea low (ASL). The ASL is a climatological low pressure system in the high-latitude Pacific Ocean whose mean position migrates from 110°W in January to 150°W in June, and its depth is greatest during austral winter (JJA) and most shallow during austral summer (DJF) (Hosking et al. 2013; Raphael et al. 2016; Turner et al. 2013). In addition, the ASL is sensitive to the climate state in the tropics and midlatitudes. Specifically, Pacific sea surface temperature variability associated with El Niño–Southern Oscillation (ENSO) impacts the southern annular mode (SAM) (L’Heureux and Thompson 2006; Fogt et al. 2011; Schneider et al. 2012) and initiates a poleward atmospheric Rossby wave response which affects the two Pacific–South American patterns (PSA1 and PSA2) (Mo and Higgins 1998; Schneider et al. 2012; Yiu and Maycock 2019). The PSA1 and PSA2 patterns are defined as the second and third EOFs of the extratropical 500-hPa geopotential height field or sea level pressure field (with SAM being the first EOF) (Kidson 1988; Yu et al. 2012) and impact atmospheric circulation throughout the mid- to high latitudes of the Southern Hemisphere (Raphael 2004; Turner 2004). Studies show that the impacts of PSA1 and PSA2 on the Amundsen Sea region are strongest during austral winter and spring, and weakest during austral summer. During austral summer (DJF), the subtropical and polar jets split and interfere with the Rossby wave train. This interference minimizes the influence of PSA1 and PSA2 and leaves SAM as the dominant mode of variability during austral summer (Bals-Elsholz et al. 2001; Jin and Kirtman 2009; Mo 2000; Schneider et al. 2012; Williams et al. 2007).
The relationships between ASL variability and the SAM, PSA1, and PSA2 are important to understanding heat and moisture transport accurately over the WAIS on inter- and intraseasonal time scales. Notably, during positive SAM, the ASL shifts eastward toward the Antarctic Peninsula and its mean depth increases initiating stronger onshore winds near the Antarctic Peninsula, the Bellingshausen Sea, and the eastern Amundsen Sea. Correspondingly, during a positive SAM, stronger offshore winds are found near the western Amundsen and eastern Ross Seas (Fogt et al. 2011; Hosking et al. 2013; Raphael et al. 2016). During positive SAM, the onshore winds bring relatively warm and moist maritime air to the eastern WAIS and reduce sea ice extent off the adjacent coast. Ding et al. (2012) show that the recent widespread warming of the WAIS during austral winter (JJA) is due to similar ASL-related intrusions of maritime air. The authors attribute the ASL changes to PSA2 variability, rather than SAM variability, which is found to be associated with central tropical Pacific sea surface temperature anomalies.
Reconstructions of past (pre-instrumental) Antarctic temperatures rely on stable oxygen and hydrogen isotope ratios in water that are found in ice cores (Dansgaard 1964). Although these records of δ18O and δD extend estimates of Antarctic surface temperature back hundreds of thousands of years, when looking at ice core records at the decadal to annual time scales, many nonlocal and non-temperature-related climatic processes impact the isotopic history at the coring sites (Casado et al. 2020; Münch and Laepple 2018; Steiger et al. 2017). Such processes are driven by atmospheric circulation variability, which can alter the source of precipitation (its location, surface temperature, and evaporative properties), the moisture transport pathway (its length scale), condensation temperature, and the rainout quantity before deposition on the Antarctic surface. To first order, poleward moisture follows moist isentropic surfaces with results showing that low-latitude sources are able to reach higher elevations on the Antarctic continent than moisture sourced from the nearby Southern Ocean (Noone 2008; Bailey et al. 2019). However, these pathways can change due to phases of ENSO and SAM (Noone and Simmonds 2002; Gregory and Noone 2008; Tsukernik and Lynch 2013) and related local sea ice anomalies (Noone and Simmonds 2004; Wang et al. 2020).
In addition to reflecting regional- to hemisphere-scale atmospheric circulation, the δ18O values in WAIS ice cores include local post-deposition processes such as snow transport by katabatic winds and water isotopic diffusion in the firn column (Fisher et al. 1985; Genthon et al. 2005). Furthermore, the WAIS ice cores are biased toward days with marine intrusions. As such, the recorded δ18O values are largely influenced by episodes of greater precipitation and warmer temperatures than the location’s annual mean (Steig et al. 1994; Noone and Simmonds 1998; Casado et al. 2020). While the above processes and biases limit the minimum temporal resolution of ice cores on the Antarctic Plateau to multidecadal time scales, signal-to-noise ratios favor annual, and possibly subannual, resolution for most of the WAIS region (Münch and Laepple 2018). Comparatively, studies show many Greenland ice cores are subannually resolved, permitting seasonal reconstructions of climate and weather patterns in the North Atlantic region (Vinther et al. 2010; Ortega et al. 2014; Zheng et al. 2018).
Here, we use continuous flow analysis (Jones et al. 2017b) to sample the WAIS Divide ice core at subannual resolution (WDCobs; information online at http://www.waisdivide.unh.edu/). The WDCobs δ18O time series is corrected with back-diffusion techniques to represent the δ18O value of the precipitation (δ18Op) at the time of deposition (Jones et al. 2017a). To determine the seasonal climate reconstruction capability of the ice core, we evaluate the seasonal mean WDCobs δ18Op values with respect to the MERRA-2 (Bosilovich et al. 2015; Gelaro et al. 2017) and utilize the Community Earth System Model version 1 with water isotopologue tracking (iCESM1) (Nusbaumer et al. 2017; Brady et al. 2019). With iCESM1, we define a pseudo–ice core time series that records the δ18Op at the time of deposition at the model grid cell where the WDCobs was drilled. We first assess the capability of the modeled pseudo–ice core δ18Op to accurately represent local precipitation and surface temperature changes as well as regional atmospheric variability. The iCESM1 analysis provides an upper limit to the seasonal, nonlocal climate reconstruction capability of the WDCobs. With the iCESM1 relationships established between the pseudo–ice core and regional climate, we evaluate the WDCobs seasonal δ18Op signal in comparison to the MERRA-2 climate. Although previous studies have utilized a combination of water isotope enabled models, modeled pseudoproxies, and paleoclimate proxies to reconstruct past climates (e.g., Steiger et al. 2017; Breil et al. 2021), to our knowledge this is this first study to assess the current capability as well as the potential of a high-resolution WAIS ice core to reveal information about past regional climate on seasonal time scales.
2. Data and methods
a. Datasets
1) iCESM1
This study uses a preindustrial run of the water isotopologue tracking version of the Community Earth System Model (iCESM1) (Brady et al. 2019), maintained by the National Center for Atmospheric Research. This model includes dynamically coupled atmosphere (CAM5), ocean (POP2), land (CLM4), sea ice (CICE4), and river runoff (RTM) components. The atmosphere and land component are on a 1.9° latitude × 2.5° longitude finite-volume grid with 30 and 10 vertical levels, respectively. The ocean and sea ice component use a ~1° rotated pole grid. The ocean has 60 vertical levels. CESM well simulates the preindustrial and present-day climate (Hurrell et al. 2013), and iCESM1 well simulates global variations in water isotopic ratios throughout the atmosphere, land, ocean, and sea ice (Brady et al. 2019).
We analyze the final 75 years of a 575-yr preindustrial simulation in which water isotopologues are tracked for the entire simulation throughout the coupled Earth system model (Singh et al. 2016; Nusbaumer et al. 2017). Initial stable O and H isotope ratios in the ocean come from the GISS interpolated ocean δ18O dataset (LeGrande and Schmidt 2006). To further evaluate the isotopic data, we tag each ocean basin in the Southern Hemisphere at 20° latitude bands and also tag the Antarctic sea ice and Antarctic land. The tagging permits a thorough analysis of the source and transport pathway of moisture falling on the West Antarctic region. Additionally, we create an iCESM1 “pseudo–ice core proxy” by generating a monthly time series of the δ18Op deposited at the grid cell where the WDCobs was cored (79°S, 112°W; Fig. 1, black star). Note, when calculating the δ18Op contribution from each tagged region to the pseudocore record, we weight by the total precipitation contribution from each tag. This pseudocore, along with other iCESM1 variable outputs, is used to determine the upper limit for reconstructing nonlocal climate variability from a single ice core.
(left) iCESM1 and (right) MERRA-2 ASL regional mean SLP and the locations of the ASL minimum central pressure for each season. Each year’s seasonal position of the ASL minimum central pressure (circles) plotted on the long-term annual mean sea level pressure (colored shading). iCESM1 data cover 75 years and MERRA-2 data cover from 1980 to 2017. The red contour marks the West Antarctic region. Gray shading is topography with contours at every 400 m above sea level. The black star marks the WAIS Divide Core location.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
2) MERRA-2
In this study, the Modern-Era Retrospective Analysis for Research and Applications version 2 (MERRA-2; Bosilovich et al. 2015) is used to represent meteorological observations from 1980 to 2017. The reanalysis uses the Goddard Earth Observing System Model, version 5 (GEOS-5) with its Atmospheric Data Assimilation System, version 5.12.4. We downloaded the atmospheric data from a 1.5° latitude × 1.5° longitude grid to closely align with the iCESM1 resolution (https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/data_access/). MERRA-2 notably improves upon its first version (MERRA) in the polar regions, including much better agreement with the 2-m air temperature station data from Antarctica (Gelaro et al. 2017).
3) WAIS divide core
We use a continuous flow analysis system (Jones et al. 2017b) to sample at subannual resolution the isotopic value of water from the West Antarctic Ice Sheet (WAIS) Divide ice core (WDCobs). Using improved back diffusion techniques (Jones et al. 2017a), we construct a WDCobs time series with 20 samples per year that represent the δ18Op at the time of deposition (see the red line in supplemental Fig. 1 in the online supplemental material). However, the precise date of each sample within a year is unknown. To address this issue, we assign approximate dates to each sample with respect to the local surface temperature seasonal cycle derived from the MERRA-2 data. Specifically, the 20 WDCobs δ18Op measurements per year were aligned such that the minimum δ18Op occurred around August and the maximum around January in accordance with the WDCobs location’s seasonal minimum and maximum surface temperature (supplemental Fig. 1, blue line). Next, the δ18Op data were grouped to create a monthly mean time series from 1980 through 2000, aligning with the start of the MERRA-2 data and the end of the WDCobs dating.
b. Methods
Along with δ18Op variability recorded at the WDC location, we also analyze 2-m temperature, total precipitation, sea level pressure, 10-m wind, sea ice, and integrated vapor transport variability in the West Antarctic region and surrounding area. The West Antarctic region is defined here as the land or ice sheet from the coastline to 83°S and from 90° to 180°W (Fig. 1, red outline).
To determine the prominent drivers of the West Antarctic region climate variability and the δ18Op variability at the WDC location, we look to inter- and intraseasonal variability in the Amundsen Sea low (ASL) (Baines and Fraedrich 1989), the southern annular mode (SAM) (Thompson and Wallace 2000), and the first and second mode of the Pacific–South America teleconnection (PSA1 and PSA2) (Kidson 1988). Following Hosking et al. (2013), the ASL is defined as the minimal sea level pressure (SLP) with respect to the mean SLP found over the ocean from 60° to 80°S and from 62° to 190°W (Fig. 1, shaded SLP area). We take the monthly time series of the ASL latitude–longitude position and depth and average to create seasonal and annual means for both iCESM1 and MERRA-2. The seasonal ASL positions are shown by the colored dots in Fig. 1 for MERRA-2 and iCESM1, respectively.
We define the SAM, PSA1, and PSA2 as the standardized leading three modes of variability from an empirical orthogonal function analysis of monthly mean sea level pressure poleward of 20°S (Fig. 2, shading) (Yu et al. 2012). SAM represents the leading mode of variability, characterized by a seesaw of atmospheric pressure between Antarctica and the Southern Hemisphere midlatitudes (Thompson and Wallace 2000). The second mode of variability, the PSA1, represents a wave train from the tropical Pacific to southern Argentina, which impacts geopotential height and wind anomalies across the South Pacific. The final mode of variability, PSA2, is similar to PSA1 apart from a 90° zonal phase lead by PSA1. Furthermore, the PSA1 pattern is similar to the Southern Hemisphere response to eastern Pacific ENSO variability, while the PSA2 pattern resembles the response to central Pacific ENSO variability (Mo and Higgins 1998; Rodrigues et al. 2015). Each monthly principal component time series is normalized to unit standard deviation and is averaged to create seasonal and annual time series for SAM, PSA1, and PSA2 for both iCESM1 and MERRA-2.
The three EOF patterns associated with (a)–(d) SAM, (e)–(h) PSA1, and (i)–(l) PSA2 and the linear regression slope of sea level pressure and 10-m wind velocity onto the normalized indices of SAM, PSA1, and PSA2 for DJF and JJA. The standardized leading three modes of variability from an empirical orthogonal function (EOF) analysis of monthly mean sea level pressure poleward of 20°S (SAM, PSA1, and PSA2) for iCESM1 and MERRA-2 (shading). (first column),(third column) The variance explained by each EOF is listed in the top right of each panel. Note that column 1 shows the same shaded EOF patterns as column 2, and column 3 shows the same as column 4 because only one set of spatial patterns results from the EOF analysis of monthly iCESM1 or MERRA-2 data. The DJF and JJA columns are distinguished in order to show the linear regression slopes of sea level pressure (yellow contours, every ±1.5 hPa σ−1) and 10-m wind velocity (vectors; m s−1 σ−1) onto the normalized, seasonal mean, indices of SAM, PSA1, and PSA2 for DJF and JJA. Sigma (σ) represents one standard deviation of the seasonal mean index, and yellow dotted (solid) contours represent a decrease (increase) in SLP as the index increases.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
This study concentrates on climate variations in the West Antarctic region due to variability in the ASL position or strength, as well as variability in SAM, PSA1, and PSA2. We evaluate the climate variations through two techniques. First, we find the mean difference of climate variables with respect to the upper minus lower quartile of the ASL position (most westward vs most eastward), ASL depth (deepest vs shallowest), and of the three indices (most positive vs most negative) (e.g., Fig. 3). Second, we also perform a linear regression analysis of sea level pressure, 10-m wind velocity, 2-m temperature, total precipitation, and δ18Op (iCESM1 only) upon the normalized indices of SAM, PSA1, and PSA2 (e.g., Fig. 5). The results are shown for the austral winter (JJA), summer (DJF), and annual mean (see the online supplemental material). Although not shown in the study, the austral spring (SON) and austral autumn (MAM) results generally resemble JJA, and therefore the annual mean results often display similar signatures as found in JJA.
Regional climate variability with respect to the ASL central pressure longitudinal position for DJF and JJA. Meridional wind (V10), 2-m temperature (T2M), total precipitation (TP), and sea ice fraction (CI) changes from the upper quartile (westward ASL location) minus the lower quartile (eastward ASL location) of the longitudinal position of the ASL minimum central pressure: (left) 75 years for iCESM1 and (right) 1980–2017 for MERRA-2. The center of the green (black) cross marks the mean location of the upper (lower)-quartile ASL position with ±1 standard deviation in the longitudinal and latitudinal directions (crosses). The contour marks a significant change in the climate variable at 95% confidence using a two-tailed t test.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
Variables are initially detrended to remove any spurious trends in the iCESM1 preindustrial model output and the anthropogenic trend in the MERRA-2 data. The full 75 years of the iCESM1 data and 1980–2017 MERRA-2 data are used to find the quartiles and determine significance of the mean quartile difference at 95% (two-sided t test). However, when MERRA-2 data are used in conjunction with the WDCobs δ18Op time series, the reanalysis data are limited to years 1980–2000 to match the length of the ice core record. Finally, we determine that the regression analysis shows a significant relationship if the p value is ≤0.05 and the relationship explains at least 15% of the variability (r2 ≥ 0.15). The latter constraint reflects this study’s goal to establish the main drivers of West Antarctic climate variability.
3. Results
a. Amundsen Sea low and the West Antarctic climate
Consistent with previous studies (e.g., Hosking et al. 2013), Fig. 3 shows that the longitudinal position of the ASL is a key driver of the West Antarctic surface climate. Figure 3 illustrates the change in multiple climate variables (rows) with respect to an intraseasonal shift in the ASL longitudinal position from its mean upper quartile position (farthest west; green cross) to its mean lower quartile position (farthest east; black cross). Notably, when the ASL shifts westward, the associated onshore winds (negative values) move west from the Peninsula toward the West Antarctic region (Figs. 3a–d). This meridional wind response transports relatively warm (Figs. 3e–h) and moist (Figs. 3i–l) maritime air to West Antarctica and reduces sea ice extent in the Amundsen Sea (Figs. 3m–p). These surface climate anomalies are consistent between iCESM1 and MERRA-2 in both the spatial extent and magnitude across DJF and JJA (Fig. 3, columns. For annual mean results, see supplemental Fig. 2).
The West Antarctic 2-m temperature and precipitation sensitivity to ASL longitudinal position are most significant in austral winter (JJA), with changes exceeding 5°C and 1.5 mm day−1, respectively (Figs. 3f,h,j,l). Comparatively, in austral summer (DJF) the 2-m temperature and precipitation change due to ASL longitudinal position is of smaller magnitude and only relatively small areas of this region show significant change (Figs. 3e,g,i,k). There is minimal temperature and precipitation change during DJF despite the ASL longitudinal position being most variable in DJF (represented by the distance between the green and black crosses in Fig. 3, column DJF vs column JJA), which leads to similar onshore wind anomalies above West Antarctica to those found in JJA (Figs. 3a–d). We attribute the limited change in temperature and precipitation in DJF, at least in part, to the reduced meridional temperature and humidity gradients between the midlatitude ocean and the West Antarctic region during DJF (not shown). Additionally, large-scale atmospheric circulations, which drive marine air intrusions into the West Antarctic region, are relatively muted during DJF (Schneider et al. 2012).
The ASL minimum central pressure (ASL depth) is another driver of the West Antarctic 2-m temperature and precipitation on seasonal to annual time scales. Figure 4 illustrates the change in multiple climate variables (rows) with respect to an intraseasonal shift in the ASL depth from its mean upper quartile value (deeper SLP value; green cross) to its mean lower quartile value (shallower SLP value; black cross). Deepening the ASL (with minimal change to the longitudinal position) strengthens and widens the associated clockwise atmospheric circulation, increasing the onshore wind near the peninsula and the adjacent Bellingshausen Sea while increasing the offshore wind near the Amundsen Sea and Ross Sea sectors (Figs. 4a–d). This strengthening of the winds increases temperature (Figs. 4e–h) and precipitation (Figs. 4i–l) near the Antarctic Peninsula and the eastern section of the West Antarctic region and decreases both variables in the western section. Both iCESM1 and MERRA-2 capture this ASL depth variability and agree on the magnitude of variable change and significance for each season and the annual mean (Fig. 4, columns). For annual mean results, see supplemental Fig. 2).
As in Fig. 3, but with respect to the ASL central pressure depth [deeper SLP (green cross) minus shallower SLP (black cross)].
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
b. SAM, PSA1, and PSA2 and the West Antarctic climate
The ASL position and depth are linked to the large-scale atmospheric variability in the Southern Hemisphere; namely, the SAM, PSA1, and PSA2 indices. Each index is associated with distinct patterns of sea level pressure and wind (Fig. 2; contours and vectors) which impact temperature, precipitation, and sea ice variability within and around the West Antarctic region on seasonal to annual time scales (e.g., Yu et al. 2012; Marshall and Thompson 2016; Irving and Simmonds 2016). We highlight these relationships in DJF and JJA for 2-m temperature and total precipitation by performing linear regression analysis between these two climate fields and the three indices (Figs. 5 and 6). For annual mean results, see supplemental Figs. 3 and 4.
The linear regression slope of 2-m temperature onto the normalized indices of (a)–(d) SAM, (e)–(h) PSA1, and (i)–(l) PSA2 for DJF and JJA. The units are °C σ−1, where σ is one standard deviation of the index. The shading is the slope with stippling where p value is ≤0.05 and r2 is ≥0.15. The center of the green (black) cross marks the mean location of the upper (lower)-quartile ASL position with respect to each index, with ±1 standard deviation in the longitudinal and latitudinal directions (crosses).
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
As in Fig. 5, but for the linear regression with total precipitation.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
Beginning with the relationship between SAM and 2-m temperature, the top row of Fig. 5 shows the regression slope (shading) with stippling where the regression slope is significant at 95% and explains at least 15% of the variability in the temperature field during DJF and JJA. These two criteria are met at the WDC location (Fig. 5, black star) during DJF and JJA in iCESM1 and met during JJA in MERRA-2 (Table 1). We attribute the strong negative slope relationship between SAM and the West Antarctic region 2-m temperature field to the SAM index being significantly correlated with ASL minimum central pressure (Pearson R, p ≤ 0.05; supplemental Table 1) in both iCESM1 and MERRA-2 across all seasons: deep anomalies occur during positive SAM and vice versa. Therefore, the pattern and magnitude of the regression slope in Fig. 5 (top row) are very similar to the 2-m temperature variability due the ASL central pressure as shown in Fig. 4 (second row). Furthermore, the regression slope pattern can be explained similarly by a change in the ASL depth: as the SAM index increases the ASL deepens, increasing the onshore winds nears the peninsula and eastern part of the West Antarctic region and increasing the offshore winds in the western part of the West Antarctic region, including at the WDC location (Figs. 2a–d). This enhanced circulation brings more cold continental air to the WDC location. This relationship is stronger during austral winter compared to austral summer for MERRA-2 (15% vs 2% variance explained), but similar between the seasons in iCESM1 (16% vs 19% variance explained; Table 1). The low 2-m temperature variance explained by SAM in DJF for the MERRA-2 data can be attributed to the relatively low wind anomalies associated with the SAM index near the WDC location (Fig. 2c).
Linear regression statistics at the WDC location of iCESM1 and MERRA-2 2-m temperature, total precipitation, and δ18Op (iCESM1 pseudo-WDC or WDCobs) regressed onto SAM, PSA1, and PSA2 during DJF and JJA. Slope values are in boldface if the p value (p-val) is ≤0.05 and the r2 value is ≥0.15. Slope values are in italics if criteria are not met. The slopes are in units of °C σ−1, mm day−1 σ−1, or ‰ σ−1, where σ is one standard deviation of the index.
The top row of Fig. 6 shows the regression slope between SAM and total precipitation during DJF and JJA. The positive slope is significant for large portions of the Antarctic Peninsula and eastern part of the West Antarctic region, particularly near the coast. In these regions, moist maritime air is brought to the coast by enhanced onshore winds from an increased SAM index. On the other hand, the western part of the West Antarctic region and near the WDC location becomes drier as the SAM index increases, although the regression slope at the WDC location does not meet the criteria of significance [Fig. 6, top row (stippling) and Table 1]. Notably, in JJA the regions of significant negative and positive slope near the WDC location from the west and the east, respectively (Figs. 6b,d), which suggests that the precipitation anomalies with respect to SAM may not reach the local high point of elevation where the WDC was cored (Fig. 1, black contours).
Similar to SAM, the PSA1 index is most strongly correlated with the depth of the ASL during JJA (supplemental Table 1). The middle row of Fig. 5 shows that as the PSA1 index increases (associated with a shallowing of the ASL), temperatures near the Antarctic Peninsula and eastern portion of the West Antarctic region cool by upward of 1.5°C σ−1 while the western portion of the West Antarctic region and Ross Sea region warm by upward of 1.5°C σ−1. The largest anomalies occur during JJA. This regression slope pattern can be explained by the following: as the PSA1 index increases, the ASL shallows, decreasing the onshore winds and maritime air brought to the Antarctic Peninsula and the eastern West Antarctic region (Figs. 2e–h). Concurrently, cold continental offshore winds in the western West Antarctic region and Ross Sea region decrease, resulting in warmer temperatures in these regions. However, the regression slope at the WDC location is only significant (and explains 23% of the variance) during JJA in MERRA-2 and is equal to −1.36°C σ−1 (Table 1). The lack of significance in DJF can be attributed, at least in part, to the strong wind anomalies associated with PSA1 being located offshore rather than near the WDC location during DJF (Figs. 2e,g).
The middle row of Fig. 6 shows that the regression slope between PSA1 and the total precipitation field depends on the season. During DJF in both iCESM1 and MERRA-2, the climatological position of the ASL is located farther east than during JJA (Fig. 6, green and black crosses). This moves the offshore component of the ASL clockwise circulation to reside over the WDC location and the western part of the West Antarctic region. Consequently, as the PSA1 index increases in DJF (associated with a shallowing of the ASL and a shift eastward) these cold and dry offshore winds decrease and permit moisture from lower latitudes to reach the WDC location through the Amundsen and Ross Sea regions resulting in the positive regression slope for iCESM1 (significant) and MERRA-2 (not significant) (Figs. 6e,g). However, during JJA the ASL position is located farther west, as compared to DJF, which brings the onshore component of the ASL clockwise circulation closer to the WDC location (Figs. 6f,h). In MERRA-2, an increase in the PSA1 index weakens the clockwise circulation and results in a significant negative regression slope for JJA (and explains 16% of the variance; Table 1). In contrast, in iCESM1, the wind anomalies or the moisture transport anomalies, near the WDC may not be large enough to result in a significant regression slope with respect to PSA1 and precipitation during JJA.
Contrary to the SAM and PSA1, the PSA2 index is not significantly correlated with ASL depth, but rather with the longitudinal position of the ASL minimum central pressure: eastward anomalies occur during positive PSA2 and vice versa (supplemental Table 1). In both austral winter and summer, this significant eastward displacement of the ASL during positive PSA2 shifts the associated onshore winds toward the Weddell Sea while the offshore winds enhance and dominate in the West Antarctic region (Figs. 2i–l). This shift in circulation causes significant decreases in temperature and precipitation with respect to an increase in PSA2 across the West Antarctic region in both DJF and JJA (Figs. 5 and 6i–l). Correspondingly, this temperature and precipitation response is very similar to the quartile difference pattern with respect to longitudinal variability as shown in Fig. 3 (second and third row); note that the sign is opposite in Fig. 3 because the ASL mean east position was subtracted from the mean west position.
The PSA2 and 2-m temperature negative regression slope is significant at the WDC location for both iCESM1 and MERRA-2 during JJA and explains 53% and 26% of the variance, respectively (Figs. 5j,l; Table 1). Consistent with the offshore wind anomalies in the West Antarctic region as the PSA2 index rises, the negative regression slope between PSA2 and precipitation is significant at the WDC location for both iCESM1 and MERRA-2 during JJA and explains 49% and 15% of the variance, respectively (Figs. 6j,l; Table 1). However, during DJF the areas with significant regression slope between PSA2 and the temperature or precipitation fields reside near the West Antarctic coast, with no significance found at the WDC location (Figs. 5 and 6i,k; Table 1). This result is consistent with the reduced circulation anomalies at the WDC location associated with PSA2 during DJF as compared to JJA for both iCESM1 and MERRA-2 (Figs. 2i–l).
To summarize, SAM, PSA1, and PSA2 each drive distinctive atmospheric conditions around Antarctica with large changes to 2-m temperature and total precipitation in the West Antarctic region, particularly during JJA (Figs. 5 and 6, shading). During JJA, the iCESM1 variations in 2-m temperature at the WDC location can be explained primarily by PSA2 (53%) with a secondary influence by SAM (16%) (Table 1). Similarly, iCESM1 variations in JJA precipitation at the WDC location can be explained primarily by PSA2 (49%), but with minimal contributions from SAM and PSA1. For MERRA-2, the JJA temperature variability at the WDC location is explained primarily by PSA2 (26%) along with contributions from both SAM (15%) and PSA1 (23%). Last, for MERRA-2 the JJA precipitation variability at the WDC location is explained by a similar contribution from PSA1 (16%) and PSA2 (15%) with minimal contribution from SAM. Therefore, the iCESM1 and MERRA-2 results agree that PSA2, and thus the longitudinal position of the ASL, explains the most variance of the temperature and precipitation record at the WDC location during JJA.
c. δ18Op and the West Antarctic climate
Moving beyond the spatial response of the two-dimensional temperature and precipitation field variables to atmospheric conditions, we set forth to understand how a pointwise record of δ18Op in West Antarctica can inform reconstructions of Southern Hemisphere atmospheric variables and circulation. The goal of such an analysis is to establish the ability of a subannually resolved ice core to inform the past regional climate state on seasonal to annual time scales. To investigate this proposal, we utilize the WDCobs record that is corrected for diffusion to represent the δ18Op at the time of deposition and is adjusted to monthly intervals (see section 2b), as well as a monthly time series of the iCESM1 δ18Op at the model grid point that best represents the WDC location (79.46°S, 112.09°W; 1806-m elevation). We refer to the latter time series as the iCESM1 pseudo-WDC δ18Op record. Both isotopic time series are linearly regressed with the 2-m temperature and total precipitation fields from iCESM1 and MERRA-2, respectively.
The top row of Fig. 7 shows the regression slope of the temperature field regressed onto the pseudo-WDC and WDCobs δ18Op time series (shading) during DJF and JJA (for annual mean results, see supplemental Fig. 5). The results show that an increase (enrichment) in δ18Op at the WDC location corresponds with warmer temperatures across much of the West Antarctic region, particularly during JJA. Notably, during JJA, the 2-m temperature at the WDC location explains 62% (iCESM1 pseudo-WDC) and 24% (WDCobs) of the variance found in the respective δ18Op time series (supplemental Table 2). The bottom row of Fig. 7 shows the regression slope of the total precipitation field regressed onto the two δ18Op time series (shading) during DJF and JJA. The results show that an increase (enrichment) in δ18Op at the WDC location corresponds with more precipitation throughout most of the West Antarctic region, particularly near the coast and during JJA. However, this positive relationship between δ18Op at the WDC location and precipitation is only significant for large portions of the West Antarctic region in iCESM1 during JJA (Figs. 7e–h, stippling). For this case, the precipitation at the WDC location explains 40% of the variance found in the pseudo-WDC δ18Op time series (supplemental Table 2).
The linear regression slope of (a)–(d) 2-m temperature and (e)–(h) total precipitation onto the iCESM1 pseudo-WDC or WDCobs δ18Op time series for DJF and JJA. The shading is the slope with stippling where p value is ≤0.05 and r2 is ≥ 0.15. The center of the green (black) cross marks the mean location of the upper (lower)-quartile ASL position with respect to either the pseudo-WDC or WDCobs δ18Op time series, with ±1 standard deviation in the longitudinal and latitudinal directions (crosses).
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
d. δ18Op and the response to SAM, PSA1, and PSA2
The above results exemplify that 2-m temperature and total precipitation influence the δ18Op at the WDC location in both the iCESM1 and WDCobs/MERRA-2 results. Therefore, it is consistent that the δ18Op field responds similarly to variability in SAM, PSA1, and PSA2 (note that only iCESM1 has a gridded δ18Op field). Figure 8 shows the iCESM1 linear regression slope of the δ18Op field onto the normalized indices of SAM, PSA1, and PSA2 during DJF and JJA (for annual mean results, see supplemental Fig. 6). At the WDC location, the regression results for iCESM1 SAM and PSA2 yield a significant negative slope during JJA (Figs. 8b,f). Furthermore, PSA2 explains 39% and SAM explains 24% of the variance in the JJA pseudo-WDC δ18Op time series (Table 1). However, PSA1 does not explain a significant amount of variance in JJA at the WDC location, and no index explains a significant amount of variance in the DJF pseudo-WDC δ18Op time series. These results are consistent with the iCESM1 2-m temperature and precipitation response to each index for each season, with the largest explained variance of δ18Op at the WDC location by PSA2, a secondary influence by SAM, and a much stronger response during JJA.
As in Fig. 5, but for the linear regression with iCESM1 δ18Op.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
The regression of the WDCobs δ18Op time series onto the three indices from MERRA-2 yield a significant negative slope for PSA1 and PSA2 during JJA (Table 1). PSA1 explains 31% and PSA2 explains 26% of the variance in the JJA WDCobs δ18Op time series. However, SAM does not explain a significant amount of variance in JJA, and no index explains a significant amount of variance in the DJF WDCobs δ18Op time series. These results are consistent with the MERRA-2 2-m temperature and precipitation response to each index for each season, with PSA1 and PSA2 explaining similar significant variance of δ18Op at the WDC location and a much stronger response during JJA.
e. Drivers of moisture transport at the iCESM1 pseudo–ice core
Although the iCESM1 and WDCobs/MERRA-2 results disagree on the relative role of SAM and PSA1 on the δ18Op record at the WDC location, the datasets do agree that PSA2 explains a significant, if not a primary, amount of δ18Op variability at the WDC location and surrounding West Antarctic region. To further explain PSA2-driven δ18Op variability, this section focuses on tracking the moisture amount and pathway from ocean tagged regions to the WDC location during JJA (for the JJA SAM and PSA1 ocean tagging results, see the online supplemental text, and supplemental Figs. 7 and 8 and Table 3).
To begin, we subtract the mean lower from the mean upper quartile of the PSA2 index to calculate the change in the overall average precipitation-weighted δ18Op signal at the WDC location. The result is a significant change of −2.85‰ (Table 2, two-tailed t test at 95%). We then select four ocean-tagged regions that consistently (across indices) contribute most to both the total and the anomalous pseudo-WDC δ18Op values to focus our analysis: Pacific 60°–80°S, Pacific 40°–60°S, Indian 40°–60°S, and Indian 20°–40°S. To help analyze the changes of each tagged region’s contribution to the pseudo-WDC δ18Op record, we also calculate integrated water vapor transport (IVT) and total precipitation originating from each of the tagged regions.
Tagged region contributions to the overall (global) change in iCESM1 integrated vapor transport (IVT), total precipitation (TP), and precipitation-weighted δ18Op with respect to PSA2 during DJF and JJA. The tagged regions sum to equal the overall amount for each variable (not exact because the contributions north of 20°S are not included in this table). Two-tailed t test with upper quartile minus lower quartile mean change at 95% confidence are in boldface text; nonsignificant changes are in italics. The values are in units of kg m−1 s−1, mm day−1, and ‰ for IVT, TP, and δ18Op, respectively (Pac = Pacific Ocean; Ind = Indian Ocean; Atl = Atlantic Ocean; AA = Antarctic).
The regions that contribute most to the overall JJA PSA2-driven −2.85‰ decrease in δ18Op at the WDC location are Pacific 40°–60°S, Pacific 60°–80°S, Indian 20°–40°S, and Indian 40°–60°S by −0.65‰, −0.51‰, −0.47‰, and −0.88‰, respectively (all are significant changes at 95% confidence and are weighted by total precipitation from the respective tag region; Table 2). Accompanying the overall depletion at WDC, Figs. 9a and 9b show a relative divergence of IVT and reduced total precipitation at the WDC and across the surrounding West Antarctic region.
Four ocean locations that contribute significantly to the overall (global) change in iCESM1 integrated vapor transport (IVT), total precipitation (TP), and precipitation-weighted δ18O (δ18Op) with respect to PSA2 during JJA. (left) IVT change (shading and vectors) from the upper quartile minus the lower quartile of the JJA PSA2 index. (center) As in the left column, but for total precipitation (shading) and 10-m winds (vectors). (right) As in the left column, but for the precipitation-weighted δ18Op (shading) and sea level pressure (gray contours, every 3 hPa). Each panel is shaded only where the quartile difference is significant at the 95% level (t test). Each panel also shows the numeric quartile difference in IVT, TP, or δ18Op at the WDC location (boldface is for a significant change at the 95% level, italics for nonsignificant), the location of the WDC (black star), the geographic domain of the tagged region (orange contour), and the mean ASL location of the upper (green crosses) and lower (black crosses) quartile position with respect to PSA2, with ±1 standard deviation in the longitudinal and latitudinal directions.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
The largest contributor to the overall decrease in iCESM1 IVT and total precipitation at the WDC location with respect to PSA2 variability comes from changes from the Pacific 40°–60°S (Table 2). Compared to negative PSA2, during positive PSA2 the ASL minimum central pressure is located eastward, which shifts the associated clockwise winds toward the Antarctic Peninsula, resulting in IVT convergence near southern South America and the surrounding ocean as the anomalous winds push the moisture around the peninsula and into the western Weddell Sea (Figs. 9d,e). These wind changes (TP column, vectors) lengthen the path of moisture transport and reduce the overall IVT from the Pacific at 40°–60°S to the WDC location. As such, the total precipitation originating from the Pacific at 40°–60°S declines at the WDC and regions located eastward of the WDC where the wind anomalies are greatest. We speculate that this lengthened path of moisture transport and decrease in precipitation from the nearby ocean source results in an isotopically light anomaly at the WDC during positive PSA2 compared to negative PSA2 (Fig. 9f).
IVT and total precipitation at the WDC location also decrease from the Indian 20°–40°S and Indian 40°–60°S tagged regions (Figs. 9j–k,m–n). Relative to negative PSA2, during positive PSA2, water vapor (not shown) and precipitation at the WDC location from the Indian 20°–40°S and Indian 40°–60°S regions are isotopically lighter, possibly due to a longer transport path resulting in more rainout along the route. Furthermore, precipitation from these regions shows only a small reduction at the WDC location during positive PSA2 while total precipitation at the WDC location is significantly reduced during positive PSA2 (Figs. 9k,n vs Fig. 9b). Therefore, the relative contribution from Indian 20°–40°S and Indian 40°–60°S regions to δ18Op at the WDC location increases during positive PSA2 compared to negative PSA2 (Figs. 9l,o vs Fig. 9c). Because these are distant sites that experience even greater rainout before reaching the WDC location during positive PSA2, the relative increase in precipitation amount from these sites amplifies the relatively low δ18Op.
The lower δ18Op from the Pacific 60°–80°S region is associated with offshore versus onshore winds when comparing positive to negative PSA2, respectively (Figs. 9g–i). The offshore wind anomalies during positive PSA2 increase the sea ice extent and decrease the surface temperature in the Pacific 60°–80°S region, potentially reducing the evaporative potential of the 18O isotope, thus leading to relatively low δ18Op. This result is consistent with Wang et al. (2020) where a decrease in sea ice extent increases moisture sourced from south of 50°S and increases the evaporative potential of the 18O isotope. The shift in the ASL minimum central pressure location may play a secondary role in the relatively low δ18Op during positive PSA2 from the Pacific 60°–80°S region.
We also note that no significant overall δ18Op change is found in DJF with respect to SAM, PSA1, or PSA2 (Table 2 and supplemental Table 3). Thus, significant changes to the overall iCESM1 pseudo-WDC δ18Op record are present during JJA, not during DJF, and are only due to PSA2 and SAM. These δ18Op changes appear to be driven primarily by shifts in atmospheric circulation, which also impacts local temperature and precipitation contributions from various ocean regions.
4. Discussion
a. iCESM1 comparison with WDCobs and MERRA-2 results
Provided that the iCESM1 results offer an upper limit to the seasonal regional climate reconstruction capability from a single ice core record, the similar expression of regional climate variability from the WDCobs time series offers encouraging results. The climate response of iCESM1 and MERRA-2 largely agree with respect to the ASL, SAM, PSA1, and PSA2 variability across the West Antarctic region (Figs. 3–6). However, establishing the relative influence of SAM, PSA1, and PSA2 at the specific WDC location yields differing results between the observation-based results (WDCobs and MERRA-2) and the iCESM1 results.
Specifically, the austral winter (JJA) correlations and regression slopes between the iCESM1 2-m temperature, total precipitation, and δ18Op at the WDC location are largely driven by PSA2 variability with a secondary influence by SAM (Table 1). Conversely, the JJA correlations and regression slopes between the MERRA-2 2-m temperature, precipitation, and WDCobs δ18Op time series show a significant influence by PSA1 along with PSA2 (Table 1). This disagreement between the relative role of SAM and PSA1 at the WDC location is likely linked to small variations between iCESM1 and MERRA-2 in the EOF analysis and the associated seasonal SLP and 10-m wind field regression slopes with respect to each index (Fig. 2). This sensitivity, particularly at the WDC location, is explored further in sections 4b and 4c. Despite disagreeing on the relative role of SAM and PSA1, both the iCESM1 and the observation-based data agree that PSA2, and thus the longitudinal position of the ASL, drives the largest changes to 2-m temperature, total precipitation, and δ18Op at the WDC location during JJA.
Establishing the relative influence of SAM, PSA1, and PSA2 on δ18Op at the WDC location for austral summer (DJF) reveal less robust results. Both the iCESM1 and the observation-based data show lower-magnitude correlations and regression slopes between the 2-m temperature, total precipitation, and δ18Op at the WDC location with respect to the three indices during DJF (Table 1). Furthermore, none of the regression slopes between δ18Op at the WDC location and the three indices are significant at 95% confidence.
b. Moisture sources and pathways
Another goal of this study is to determine how the atmospheric circulation, moisture transport pathways, and sources of precipitation influence the δ18Op variability across the West Antarctic region and at the WDC location. Here, we focus on the relatively large changes in these variables that occur in JJA due to SAM, PSA1, and PSA2, and compare iCESM1 to the observation-based results. The patterns of total precipitation with respect to SAM, PSA1, and PSA2 variability show similar responses in both iCESM1 and MERRA-2 in the West Antarctic region [Fig. 10, total precipitation (TP) columns]. However, very small differences between iCESM1 and MERRA-2 in atmospheric circulation and moisture transport with respect to the indices have a large impact at the WDC location.
The overall (global) change in iCESM1 and MERRA-2 integrated vapor transport, total precipitation, and δ18Op with respect to (a)–(f) SAM, (g)–(l) PSA1, and (m)–(r) PSA2 during JJA. Same caption with respect to each IVT, TP, and δ18Op column as in Fig. 9. There is no δ18Op field for MERRA-2.
Citation: Journal of Climate 34, 24; 10.1175/JCLI-D-20-0822.1
The upper minus lower quartile difference in δ18Op at the WDC location with respect to SAM during JJA is −2.07‰ (significant) and −0.15‰ (not significant) for the iCESM1 pseudo-WDC and WDCobs, respectively (Figs. 10c,f, and supplemental Table 4). Although patterns of IVT and total precipitation change are similar, it is evident that these changes are more often significant around the West Antarctic region in iCESM1 (Fig. 10, top row). One driver of this difference between iCESM1 and MERRA-2 is the deeper mean ASL during positive SAM for iCESM1 (by 9.5 hPa) compared to MERRA-2 (by 8.8 hPa) and the coinciding increase in clockwise winds (Figs. 10b,e, vectors). The greater atmospheric response of iCESM1 to JJA SAM variability results in fewer, or weaker, warm maritime air intrusions into the West Antarctic region during positive SAM as compared to MERRA-2. At the WDC location, this appears to drive a significant depletion of δ18Op in iCESM1 but not in the WDCobs record.
The upper minus lower quartile difference in δ18Op at the WDC location with respect to PSA1 during JJA is 0.39‰ (not significant) and −1.64‰ (not significant at 95%, but p value is <0.10) for the iCESM1 pseudo-WDC and WDCobs, respectively (Figs. 10i,l, and supplemental Table 4). Again, the broad patterns of IVT and total precipitation change are similar in spatial extent and magnitude, although the ASL depth increase with respect to the PSA1 quartile difference is larger for MERRA-2 (by 5.3 hPa) compared to iCESM1 (by 3.7 hPa) and these SLP anomalies are located farther northwest of WDC location (Fig. 10i compared to Fig. 10l, contours). The consequence of the ASL and associated counterclockwise wind anomalies being located farther from the WDC in MERRA-2 likely has a large impact on the magnitude and sign of the WDCobs δ18Op anomalies. In MERRA-2, the anomalous clockwise winds are located farther from the WDC location and therefore do not bring as much moisture that has relatively high δ18O from local ocean sources. Additionally, the JJA 2-m temperature response to PSA1 shows a cooling in MERRA-2 but a warming in iCESM1 at the WDC location (Figs. 5f,h). This absence of locally sourced moisture and decrease in 2-m temperature with respect to a PSA1 increase could be responsible for the relatively large decrease of the δ18Op in the WDCobs record (−1.64‰) not seen in the iCESM1 results.
Last, the upper minus lower quartile difference in δ18Op at the WDC location with respect to PSA2 during JJA is −2.85‰ (significant) and −1.80‰ (not significant at 95%, but p value is <0.10) for the iCESM1 pseudo-WDC and the WDCobs, respectively (Figs. 10o,r, and supplemental Table 4). The eastward shift of the ASL during positive PSA2 creates offshore wind anomalies in the West Antarctic and reduced IVT and total precipitation in the West Antarctic region for both iCESM1 and MERRA-2 (Fig. 10, bottom row). To first order, the magnitude of change in these variables is larger in iCESM1 due to broader field of negative SLP anomalies across West Antarctic and the Peninsula (Figs. 10o,r, contours). Given that nearly all the iCESM1 ocean tag regions contribute depletion anomalies at the WDC location (Table 2), we suggest that the switch from onshore to offshore winds due to the eastward migration of the ASL during positive PSA2 greatly reduces marine intrusions from the midlatitude Pacific. Furthermore, we suggest that during positive PSA2 the IVT routes moisture across the Bellingshausen Sea and Peninsula before arriving depleted at the WDC location. These JJA PSA2-driven changes appear to significantly impact the δ18Op at the WDC location, resulting in the largest fluctuations of the isotopic composition of the precipitation found in the study.
c. Sensitivity of the results due to the WDC location
Throughout this study, we find that small variations in the location or depth of the ASL in response to the SAM, PSA1, or PSA2 in JJA can determine whether the pseudo-WDC or WDCobs reflects a significant δ18Op increase or decrease. Contributing to the sensitivity of the δ18Op is the location of the ice core site. The ice core was taken from the WAIS Divide, which is a ridge of local high topography (about 1800 m above sea level) that separates ice flowing to the Ross Sea in the west and to the Pine Island Bay in the east (Fig. 1, black contours). Multiple variable responses with respect to the ASL depth, SAM, and PSA1 show an abrupt change in magnitude or significance on either side of the WDC location (Figs. 4–6). Therefore, climate responses due to the ASL depth, SAM, and PSA1 indices are sensitive to small changes in the strength and position of the circulation anomalies which determine whether heat or moisture is brought to the local high point of topography. On the other hand, variable responses to the ASL longitudinal position and PSA2 are consistent across the full West Antarctic region and are thus not as sensitive to small changes in the strength or position of the ASL (Figs. 3, 5, and 6). As a result, the West Antarctic climate in iCESM1 and MERRA-2 agree best with respect to ASL longitudinal variability as demonstrated by the PSA2 index.
Furthermore, the correlations, regression slopes, and variable quartile differences with respect to ASL, SAM, PSA1, and PSA2 are larger during JJA as compared to DJF. This is partially due to the anomalous wind field associated with each EOF being located farther from the WDC location or weaker during DJF than JJA. Additionally, this difference between DJF and JJA can be attributed to the majority of maritime intrusions into the West Antarctic region occurring in non-DJF months when the polar front jet is relatively weak and permits tropically sourced Rossby waves to impact the ASL region (Nicolas and Bromwich 2011). The large intraseasonal variability of the WDCobs δ18Op in JJA helps establish significant relationships with regional climate changes due to ASL and the three indices. On the other hand, this temporal characteristic of maritime intrusions leads to a bias toward capturing days in the WDCobs record with greater than 5 mm of precipitation and surface temperatures around 10°C warmer than the mean in the nonsummer months (Noone and Simmonds 1998). Therefore, using the continuous-flow method to subannually resolve the WDCobs will result in samples that are bias toward nonsummer months in both sample size, as well as in the magnitude of the δ18O values due to diffusion.
5. Conclusions
Utilizing iCESM1 provides an upper limit to the capability of reconstructing regional climate from a single ice core δ18O record at seasonal-to-annual time scales. The “pseudo-WDC” δ18Op time series is responsive to large-scale patterns of atmospheric circulation variability, notably the SAM, PSA1, and PSA2, which drive consistent changes to the position and depth of the ASL. Furthermore, each pattern is associated with a unique combination of Southern Hemisphere moisture sources and vapor transport pathways resulting in distinct δ18Op patterns across the West Antarctic region. The iCESM1 results show that the primary influence of δ18Op variability at the WDC location is PSA2 with a secondary influence by SAM during austral winter (JJA). The atmospheric response to the ASL, SAM, PSA1, and PSA2 variability in the West Antarctic region and surrounding high latitudes is similar to responses found in the MERRA-2 data. However, at the WDC location specifically, the variability in the JJA WDCobs δ18Op time series more strongly responds to PSA1, in addition to PSA2. We attribute this disagreement among iCESM1 and MERRA-2 concerning the relative roles of SAM and PSA1, at least in part, to small differences in the associated circulation patterns that manifest into relatively large differences in climate variability at the WDC location associated with these indices.
We find that the disagreement in the relative role of SAM and PSA1 at the WDC location between iCESM1 and the observation-based results may be attributed to the location of the core site atop the WAIS Divide. Significant changes to temperature, precipitation, and δ18Op due to SAM and PSA1 variability can be found throughout the West Antarctic region, but often these anomalies do not reach the local high point of topography of the WDC location. Relatively small variations in the ASL position and depth as well as the associated winds with respect to SAM and PSA1 can determine whether the δ18Op time series at the WDC location reflects changes to these indices. In contrast, PSA2 drives a relatively large shift in the ASL longitudinal position creating significant changes to temperature, precipitation, and δ18Op throughout the West Antarctic region and at the WDC location.
To improve future studies, the addition of high-resolution cores downslope from the WDC to the east and west would help distinguish SAM and PSA1 variability, given the abrupt change in magnitude on either side of the WDC location with respect to these indices. Moreover, an increase in high-resolution ice cores in the West Antarctic region would help maximize the signal-to-noise ratio. Meanwhile, finer model resolution could potentially help distinguish δ18Op variability atop the high topography at the WAIS Divide. These additions will better establish distinct δ18O patterns across the West Antarctic Ice Sheet with respect to regional modes of climate variance—the SAM, PSA1, and PSA2. This improved understanding could be directly applied to projecting West Antarctic warming and precipitation trends due to the increasing frequency of the central Pacific El Niño conditions which are associated with the PSA2 pattern (Lee and McPhaden 2010; Ding et al. 2012; Liu et al. 2017).
Acknowledgments
C. R. Tabor acknowledges funding from the National Center for Atmosphere Research Advanced Study Program postdoctoral fellowship. The CESM project is supported primarily by the National Science Foundation (NSF). This material is based upon work supported by the National Center for Atmospheric Research, which is a major facility sponsored by the NSF under Cooperative Agreement 1852977. Computing and data storage resources, including the Cheyenne supercomputer (https://doi.org/10.5065/D6RX99HX), were provided by the Computational and Information Systems Laboratory (CISL) at NCAR. We thank two anonymous reviewers for useful comments that greatly improved this paper.
Data availability statement
Code used for data collection and analysis can be found at https://github.com/pgoddard08/iCESM1_WDC_MERRA2_Project. Please contact the corresponding author (P. B. Goddard) for assistance.
REFERENCES
Bailey, A., H. K. A. Singh, and J. Nusbaumer, 2019: Evaluating a moist isentropic framework for poleward moisture transport: Implications for water isotopes over Antarctica. Geophys. Res. Lett., 46, 7819–7827, https://doi.org/10.1029/2019GL082965.
Baines, P. G., and K. Fraedrich, 1989: Topographic effects on the mean tropospheric flow patterns around Antarctica. J. Atmos. Sci., 46, 3401–3415, https://doi.org/10.1175/1520-0469(1989)046<3401:TEOTMT>2.0.CO;2.
Bals-Elsholz, T. M., E. H. Atallah, L. F. Bosart, T. A. Wasula, M. J. Cempa, and A. R. Lupo, 2001: The wintertime Southern Hemisphere split jet: Structure, variability, and evolution. J. Climate, 14, 4191–4215, https://doi.org/10.1175/1520-0442(2001)014<4191:TWSHSJ>2.0.CO;2.
Bamber, J. L., M. Oppenheimer, R. E. Kopp, W. P. Aspinall, and R. M. Cooke, 2019: Ice sheet contributions to future sea-level rise from structured expert judgment. Proc. Natl. Acad. Sci. USA, 116, 11 195–11 200, https://doi.org/10.1073/pnas.1817205116.
Bosilovich, M. G., and Coauthors, 2015: MERRA-2 : Initial evaluation of the climate. NASA Tech. Memo. NASA/TM-2015-104606, Vol. 43, 145 pp., https://gmao.gsfc.nasa.gov/pubs/docs/Bosilovich803.pdf.
Brady, E., and Coauthors, 2019: The connected isotopic water cycle in the Community Earth System Model version 1. J. Adv. Model. Earth Syst., 11, 2547–2566, https://doi.org/10.1029/2019MS001663.
Breil, M., E. Christner, A. Cauquoin, M. Werner, M. Karremann, and G. Schädler, 2021: Applying an isotope-enabled regional climate model over the Greenland ice sheet: Effect of spatial resolution on model bias. Climate Past, 17, 1685–1699, https://doi.org/10.5194/CP-17-1685-2021.
Casado, M., T. Münch, and T. Laepple, 2020: Climatic information archived in ice cores: Impact of intermittency and diffusion on the recorded isotopic signal in Antarctica. Climate Past, 16, 1581–1598, https://doi.org/10.5194/cp-16-1581-2020.
Dansgaard, W., 1964: Stable isotopes in precipitation. Tellus, 16, 436–468, https://doi.org/10.3402/tellusa.v16i4.8993.
DeConto, R. M., and Coauthors, 2021: The Paris climate agreement and future sea-level rise from Antarctica. Nature, 593, 83–89, https://doi.org/10.1038/s41586-021-03427-0.
Ding, Q., E. J. Steig, D. S. Battisti, and J. M. Wallace, 2012: Influence of the tropics on the southern annular mode. J. Climate, 25, 6330–6348, https://doi.org/10.1175/JCLI-D-11-00523.1.
Edwards, T. L., and Coauthors, 2021: Projected land ice contributions to twenty-first-century sea level rise. Nature, 593, 74–82, https://doi.org/10.1038/s41586-021-03302-y.
Fisher, D. A., N. Reeh, and H. B. Clausen, 1985: Stratigraphic noise in time series derived from ice cores. Ann. Glaciol., 7, 76–83, https://doi.org/10.3189/S0260305500005942.
Fogt, R. L., D. H. Bromwich, and K. M. Hines, 2011: Understanding the SAM influence on the South Pacific ENSO teleconnection. Climate Dyn., 36, 1555–1576, https://doi.org/10.1007/s00382-010-0905-0.
Frieler, K., and Coauthors, 2015: Consistent evidence of increasing Antarctic accumulation with warming. Nat. Climate Change, 5, 348–352, https://doi.org/10.1038/nclimate2574.
Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1.
Genthon, C., S. Kaspari, and P. A. Mayewski, 2005: Interannual variability of the surface mass balance of West Antarctica from ITASE cores and ERA-40 reanalyses, 1958–2000. Climate Dyn., 24, 759–770, https://doi.org/10.1007/s00382-005-0019-2.
Goddard, P. B., C. O. Dufour, J. Yin, S. M. Griffies, and M. Winton, 2017: CO2-induced ocean warming of the Antarctic continental shelf in an eddying global climate model. J. Geophys. Res. Oceans, 122, 8079–8101, https://doi.org/10.1002/2017JC012849.
Gregory, S., and D. Noone, 2008: Variability in the teleconnection between the El Niño–Southern Oscillation and West Antarctic climate deduced from West Antarctic ice core isotope records. J. Geophys. Res., 113, D17110, https://doi.org/10.1029/2007JD009107.
Hosking, J. S., A. Orr, G. J. Marshall, J. Turner, and T. Phillips, 2013: The influence of the Amundsen–Bellingshausen Seas low on the climate of West Antarctica and its representation in coupled climate model simulations. J. Climate, 26, 6633–6648, https://doi.org/10.1175/JCLI-D-12-00813.1.
Hurrell, J. W., and Coauthors, 2013: The Community Earth System Model: A framework for collaborative research. Bull. Amer. Meteor. Soc., 94, 1339–1360, https://doi.org/10.1175/BAMS-D-12-00121.1.
Irving, D., and I. Simmonds, 2016: A new method for identifying the Pacific–South American pattern and its influence on regional climate variability. J. Climate, 29, 6109–6125, https://doi.org/10.1175/JCLI-D-15-0843.1.
Jenkins, A., P. Dutrieux, S. Jacobs, E. J. Steig, G. H. Gudmundsson, J. Smith, and K. J. Heywood, 2016: Decadal ocean forcing and Antarctic ice sheet response: Lessons from the Amundsen Sea. Oceanography, 29, 106–117, https://doi.org/10.5670/oceanog.2016.103.
Jin, D., and B. P. Kirtman, 2009: Why the Southern Hemisphere ENSO responses lead ENSO. J. Geophys. Res., 114, D23101, https://doi.org/10.1029/2009JD012657.
Jones, T. R., K. M. Cuffey, J. W. C. White, E. J. Steig, C. Buizert, B. R. Markle, J. R. McConnell, and M. Sigl, 2017a: Water isotope diffusion in the WAIS divide ice core during the Holocene and last glacial. J. Geophys. Res. Earth Surf., 122, 290–309, https://doi.org/10.1002/2016JF003938.
Jones, T. R., J. W. C. White, E. J. Steig, B. H. Vaughn, V. Morris, V. Gkinis, B. R. Markle, and S. W. Schoenemann, 2017b: Improved methodologies for continuous-flow analysis of stable water isotopes in ice cores. Atmos. Meas. Tech., 10, 617–632, https://doi.org/10.5194/amt-10-617-2017.
Kidson, J. W., 1988: Interannual variations in the Southern Hemisphere circulation. J. Climate, 1, 1177–1198, https://doi.org/10.1175/1520-0442(1988)001<1177:IVITSH>2.0.CO;2.
Lee, T., and M. J. McPhaden, 2010: Increasing intensity of El Niño in the central-equatorial Pacific. Geophys. Res. Lett., 37, L14603, https://doi.org/10.1029/2010GL044007.
LeGrande, A. N., and G. A. Schmidt, 2006: Global gridded data set of the oxygen isotopic composition in seawater. Geophys. Res. Lett., 33, L12604, https://doi.org/10.1029/2006GL026011.
Lenaerts, J. T. M., M. Vizcaino, J. Fyke, L. van Kampenhout, and M. R. van den Broeke, 2016: Present-day and future Antarctic ice sheet climate and surface mass balance in the Community Earth System Model. Climate Dyn., 47, 1367–1381, https://doi.org/10.1007/s00382-015-2907-4.
L’Heureux, M. L., and D. W. J. Thompson, 2006: Observed relationships between the El Niño–Southern Oscillation and the extratropical zonal-mean circulation. J. Climate, 19, 276–287, https://doi.org/10.1175/JCLI3617.1.
Liu, Y., and Coauthors, 2017: Recent enhancement of Central Pacific El Niño variability relative to last eight centuries. Nat. Commun., 8, 15386, https://doi.org/10.1038/ncomms15386.
Marshall, G. J., and D. W. J. Thompson, 2016: The signatures of large-scale patterns of atmospheric variability in Antarctic surface temperatures. J. Geophys. Res. Atmos., 121, 3276–3289, https://doi.org/10.1002/2015JD024665.
Mo, K. C., 2000: Relationships between low-frequency variability in the Southern Hemisphere and sea surface temperature anomalies. J. Climate, 13, 3599–3610, https://doi.org/10.1175/1520-0442(2000)013<3599:RBLFVI>2.0.CO;2.
Mo, K. C., and R. W. Higgins, 1998: The Pacific–South American modes and tropical convection during the Southern Hemisphere winter. Mon. Wea. Rev., 126, 1581–1596, https://doi.org/10.1175/1520-0493(1998)126<1581:TPSAMA>2.0.CO;2.
Münch, T., and T. Laepple, 2018: What climate signal is contained in decadal- to centennial-scale isotope variations from Antarctic ice cores? Climate Past, 14, 2053–2070, https://doi.org/10.5194/cp-14-2053-2018.
Nicolas, J. P., and D. H. Bromwich, 2011: Climate of West Antarctica and influence of marine air intrusions. J. Climate, 24, 49–67, https://doi.org/10.1175/2010JCLI3522.1.
Noone, D., 2008: The influence of midlatitude and tropical overturning circulation on the isotopic composition of atmospheric water vapor and Antarctic precipitation. J. Geophys. Res., 113, D04102, https://doi.org/10.1029/2007JD008892.
Noone, D., and I. Simmonds, 1998: Implications for the interpretation of ice-core isotope data from analysis of modelled Antarctic precipitation. Ann. Glaciol., 27, 398–402, https://doi.org/10.3189/1998AoG27-1-398-402.
Noone, D., and I. Simmonds, 2002: Annular variations in moisture transport mechanisms and the abundance of δ18O in Antarctic snow. J. Geophys. Res., 107, 4742, https://doi.org/10.1029/2002JD002262.
Noone, D., and I. Simmonds, 2004: Sea ice control of water isotope transport to Antarctica and implications for ice core interpretation. J. Geophys. Res., 109, D07105, https://doi.org/10.1029/2003JD004228.
Nusbaumer, J., T. E. Wong, C. Bardeen, and D. Noone, 2017: Evaluating hydrological processes in the Community Atmosphere Model version 5 (CAM5) using stable isotope ratios of water. J. Adv. Model. Earth Syst., 9, 949–977, https://doi.org/10.1002/2016MS000839.
Ortega, P., D. Swingedouw, V. Masson-Delmotte, C. Risi, B. Vinther, P. Yiou, R. Vautard, and K. Yoshimura, 2014: Characterizing atmospheric circulation signals in Greenland ice cores: Insights from a weather regime approach. Climate Dyn., 43, 2585–2605, https://doi.org/10.1007/s00382-014-2074-z.
Paolo, F. S., L. Padman, H. A. Fricker, S. Adusumilli, S. Howard, and M. R. Siegfried, 2018: Response of Pacific-sector Antarctic ice shelves to the El Niño/Southern Oscillation. Nat. Geosci., 11, 121–126, https://doi.org/10.1038/s41561-017-0033-0.
Raphael, M. N., 2004: A zonal wave 3 index for the Southern Hemisphere. Geophys. Res. Lett., 31, https://doi.org/10.1029/2004GL020365.
Raphael, M. N., and Coauthors, 2016: The Amundsen Sea low: Variability, change, and impact on Antarctic climate. Bull. Amer. Meteor. Soc., 97, 111–121, https://doi.org/10.1175/BAMS-D-14-00018.1.
Rignot, E., and S. S. Jacobs, 2002: Rapid bottom melting widespread near Antarctic ice sheet grounding lines. Science, 296, 2020–2023, https://doi.org/10.1126/science.1070942.
Rodrigues, R. R., E. J. D. Campos, and R. Haarsma, 2015: The impact of ENSO on the South Atlantic subtropical dipole mode. J. Climate, 28, 2691–2705, https://doi.org/10.1175/JCLI-D-14-00483.1.
Schneider, D. P., Y. Okumura, and C. Deser, 2012: Observed Antarctic interannual climate variability and tropical linkages. J. Climate, 25, 4048–4066, https://doi.org/10.1175/JCLI-D-11-00273.1.
Shepherd, A., and Coauthors, 2018: Mass balance of the Antarctic ice sheet from 1992 to 2017. Nature, 558, 219–222, https://doi.org/10.1038/s41586-018-0179-y.
Singh, H. K. A., C. M. Bitz, A. Donohoe, J. Nusbaumer, and D. C. Noone, 2016: A mathematical framework for analysis of water tracers. Part II: Understanding large-scale perturbations in the hydrological cycle due to CO2 doubling. J. Climate, 29, 6765–6782, https://doi.org/10.1175/JCLI-D-16-0293.1.
Smith, B., and Coauthors, 2020: Pervasive ice sheet mass loss reflects competing ocean and atmosphere processes. Science, 368, 1239–1242, https://doi.org/10.1126/science.aaz5845.
Steig, E. J., P. M. Grootes, and M. Stuiver, 1994: Seasonal precipitation timing and ice core records. Science, 266, 1885–1886, https://doi.org/10.1126/science.266.5192.1885.
Steiger, N. J., E. J. Steig, S. G. Dee, G. H. Roe, and G. J. Hakim, 2017: Climate reconstruction using data assimilation of water isotope ratios from ice cores. J. Geophys. Res., 122, 1545–1568, https://doi.org/10.1002/2016JD026011.
Thompson, D. W. J., and J. M. Wallace, 2000: Annular modes in the extratropical circulation. Part II: Trends. J. Climate, 13, 1018–1036, https://doi.org/10.1175/1520-0442(2000)013<1018:AMITEC>2.0.CO;2.
Tsukernik, M., and A. H. Lynch, 2013: Atmospheric meridional moisture flux over the Southern Ocean: A story of the Amundsen Sea. J. Climate, 26, 8055–8064, https://doi.org/10.1175/JCLI-D-12-00381.1.
Turner, J., 2004: The El Niño–southern oscillation and Antarctica. Int. J. Climatol., 24 (1), 1–31, https://doi.org/10.1002/joc.965.
Turner, J., T. J. Bracegirdle, T. Phillips, G. J. Marshall, and J. S. Hosking, 2013: An initial assessment of Antarctic Sea ice extent in the CMIP5 models. J. Climate, 26, 1473–1484, https://doi.org/10.1175/JCLI-D-12-00068.1.
Vinther, B. M., P. D. Jones, K. R. Briffa, H. B. Clausen, K. K. Andersen, D. Dahl-Jensen, and S. J. Johnsen, 2010: Climatic signals in multiple highly resolved stable isotope records from Greenland. Quat. Sci. Rev., 29, 522–538, https://doi.org/10.1016/j.quascirev.2009.11.002.
Wang, H., J. G. Fyke, J. T. M. Lenaerts, J. M. Nusbaumer, H. Singh, D. Noone, P. J. Rasch, and R. Zhang, 2020: Influence of sea-ice anomalies on Antarctic precipitation using source attribution in the Community Earth System Model. Cryosphere, 14, 429–444, https://doi.org/10.5194/tc-14-429-2020.
Williams, L. N., S. Lee, and S. W. Son, 2007: Dynamics of the Southern Hemisphere spiral jet. J. Atmos. Sci., 64, 548–563, https://doi.org/10.1175/JAS3939.1.
Yiu, Y. Y. S., and A. C. Maycock, 2019: On the seasonality of the El Niño teleconnection to the Amundsen Sea region. J. Climate, 32, 4829–4845, https://doi.org/10.1175/JCLI-D-18-0813.1.
Yu, L., Z. Zhang, M. Zhou, S. Zhong, D. Lenschow, H. Hsu, H. Wu, and B. Sun, 2012: Influence of the Antarctic oscillation, the Pacific–South American modes and the El Niño–Southern Oscillation on the Antarctic surface temperature and pressure variations. Antarct. Sci., 24, 59–76, https://doi.org/10.1017/S095410201100054X.
Zheng, M., J. Sjolte, F. Adolphi, B. Møllesøe Vinther, H. C. Steen-Larsen, T. J. Popp, and R. Muscheler, 2018: Climate information preserved in seasonal water isotope at NEEM: Relationships with temperature, circulation and sea ice. Climate Past, 14, 1067–1078, https://doi.org/10.5194/cp-14-1067-2018.
